Secondary Logo

Journal Logo

Bispectral Index Can Reliably Detect Deep Sedation in Mechanically Ventilated Patients: A Prospective Multicenter Validation Study

Wang, Zhu-Heng MD*†; Chen, Han MD*‡; Yang, Yan-Lin MB; Shi, Zhong-Hua MD*; Guo, Qing-Hua MB; Li, Yu-Wei MB; Sun, Li-Ping BS; Qiao, Wei BS; Zhou, Guan-Hua MD; Yu, Rong-Guo MD; Yin, Kai MD§; He, Xuan BS*; Xu, Ming BS*; Brochard, Laurent J. MD‖¶; Zhou, Jian-Xin MD*

doi: 10.1213/ANE.0000000000001786
Critical Care and Resuscitation: Original Clinical Research Report
Free
SDC

BACKGROUND: Excessively deep sedation is prevalent in mechanically ventilated patients and often considered suboptimal. We hypothesized that the bispectral index (BIS), a quantified electroencephalogram instrument, would accurately detect deep levels of sedation.

METHODS: We prospectively enrolled 90 critically ill mechanically ventilated patients who were receiving sedation. The BIS was monitored for 24 hours and compared with the Richmond Agitation Sedation Scale (RASS) evaluated every 4 hours. Deep sedation was defined as a RASS of −3 to −5. Threshold values of baseline BIS (the lowest value before RASS assessment) and stimulated BIS (the highest value after standardized assessment) for detecting deep sedation were determined in a training set (45 patients, 262 RASS assessments). Diagnostic accuracy was then analyzed in a validation set (45 patients, 264 RASS assessments).

RESULTS: Deep sedation was only prescribed in 6 (6.7%) patients, but 76 patients (84.4%) had at least 1 episode of deep sedation. Thresholds for detecting deep sedation of 50 for baseline and 80 for stimulated BIS were identified, with respective areas under the receiver-operating characteristic curve of 0.771 (95% confidence interval, 0.714–0.828) and 0.805 (0.752–0.857). The sensitivity and specificity of baseline BIS were 94.0% and 66.5% and of stimulated BIS were 91.0% and 66.5%. When baseline and stimulated BIS were combined, the sensitivity, specificity, and clinical utility index were 85.0% (76.1%–91.1%), 85.9% (79.5%–90.7%), and 66.9% (57.8%–76.0%), respectively.

CONCLUSIONS: Combining baseline and stimulated BIS may help detect deep sedation in mechanically ventilated patients.

Published ahead of print December 22, 2016.

From the *Department of Critical Care Medicine, Beijing Tiantan Hospital, and Department of Critical Care Medicine, Daxing Teaching Hospital, Capital Medical University, Beijing, China; Surgical Intensive Care Unit, Fujian Provincial Clinical College Hospital, Fujian Medical University, Fuzhou, Fujian, China; §Intensive Care Unit, Beijing Electric Power Hospital, Capital Medical University, Beijing, China; Keenan Research Centre, St Michael’s Hospital, Toronto, Canada; and Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Canada.

Published ahead of print December 22, 2016.

Accepted for publication November 2, 2016.

Funding: The study was supported by grants from the Beijing Municipal Administration of Hospital (ZYLX201502, DFL20150502). The sponsor had no role in the study design, data collection, data analysis, data interpretation, or writing of the report.

The authors declare no conflicts of interests.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website.

Address correspondence to Jian-Xin Zhou, MD, Department of Critical Care Medicine, Beijing Tiantan Hospital, Capital Medical University, No 6, Tiantan Xili, Dongcheng District, Beijing 100050, China. Address e-mail to zhoujx.cn@icloud.com.

Observational studies find that the prevalence of deep sedation ranges from 35% to 68% in mechanically ventilated patients and that deep levels of sedation are associated with adverse clinical outcomes.1–4 Randomized controlled trials and quality improvement projects additionally suggest that targeting light sedation levels reduces sedative usage and shortens the duration of mechanical ventilation and intensive care unit (ICU) length of stay. Recent revised clinical guidelines also recommend that deep sedation should be avoided in patients in the ICU.5

Successfully maintaining a light level of sedation in clinical practice, however, is difficult.6–8 Patient arousal, sleep cycles, noise, and clinical care activities may all cause sedation levels to fluctuate. An important component in maintaining optimal sedation is the ongoing assessment of sedation depth, which primarily is evaluated with the use of subjective scales.5,9 Although these scales provide a standardized assessment of sedation levels, their intermittent nature hinders their ability to detect undesired deep sedation.10 The bispectral index (BIS), a quantified electroencephalogram instrument,11 has been used for continuous, objective evaluation of sedation depth in mechanically ventilated patients.12,13

Most studies investigating the correlation between BIS values and commonly used sedation scales have yielded conflicting results.14–22 This uncertainty may be partly because of confounders in the ICU environment and variability in the ICU patient population11–13,23; however, most studies have targeted a specific BIS range,24–27 and confounders that falsely increase the BIS value may result in unnecessary deep sedation. Another potential use of BIS monitoring to assess ICU sedation depth might rather be to act as a screening tool to detect and avoid undesired deep sedation. During continuous BIS monitoring, a low value could alert the ICU caregiver to evaluate the patient for unintended deep sedation. In this way, continuous BIS monitoring may reduce caregiver workload in resource-poor ICUs.

In the present study, we prospectively enrolled adult patients who were mechanically ventilated and sedated. The BIS was monitored continuously, and sedation depth was evaluated intermittently by use of the Richmond Agitation Sedation Scale (RASS). The primary aim was to assess the diagnostic accuracy of BIS in detecting deep sedation when compared with the RASS scale. We hypothesized that a low baseline BIS and inadequate response of BIS to external stimulation could detect deep sedation.

Back to Top | Article Outline

METHODS

Study Setting and Population

The present study was performed in 4 ICUs of academic hospitals in China: Daxing Teaching Hospital (general ICU), Beijing Tiantan Hospital (general ICU), Beijing Electric Power Hospital (general ICU), and Fujian Provincial Clinical College Hospital (surgical ICU). The local institutional review board of each participating hospital approved this study (KY2014-05-013, KY2015-02-011, 20150215CC, KY2014-05-013, and KY-2015-CC-001, respectively). The study was registered at ClinicalTrials.gov (NCT02203344 in July 28, 2014, and NCT02439840 in May 7, 2015, by Principal Investigator Jian-Xin Zhou). Written informed consents were obtained from patients or appropriate substitute decision makers.

Study design, performance, and report were complied with the Standards for Reporting of Diagnostic Accuracy guidelines. Adult patients were eligible for inclusion if they were intubated and ventilated within the previous 24 hours, were receiving intravenous sedatives and/or analgesics, and were expected to require mechanical ventilation for longer than 24 hours. The exclusion criteria included (1) age <18 or >65 years; (2) concurrent use of muscle relaxants; (3) diagnosed or suspected brain diseases, including brain trauma, intracranial hemorrhage, stroke, brain tumors, hypoxic-ischemic encephalopathy, metabolic encephalopathy, epilepsy, and meningitis; and (4) diagnosed conditions that might impair consciousness, including hypoxemia with partial pressure of oxygen in arterial blood less than 60 mm Hg, hypotension with systolic blood pressure less than 90 mm Hg, hypoglycemia with blood glucose concentration less than 4.1 mmol/L, anemia with hemoglobin concentration less than 70 g/L, and body temperature below 36°C.

Back to Top | Article Outline

Study Procedure

During the study, no attempts were made to change or influence routine patient care. The level of sedation was evaluated by RASS every 4 hours or as needed.28 The choice of sedation strategy was at the discretion of the physician on duty and included intermittent or continuous intravenous infusion of propofol or midazolam or continuous intravenous infusions of dexmedetomidine. Fentanyl was administered when the patient complained of pain. The level of sedation was titrated to a RASS of −2 to +1 per ICU routine, unless the managing physician prescribed a deeper sedation level (a RASS ≤ −3). The main reason for deep sedation in mechanical ventilated patients without brain injury was to treat tachypnea.

After enrollment, BIS was monitored continuously for 24 hours with a BIS A-2000 XP monitor (Medtronic, Dublin, Ireland: Host version 3.31) with a Quattro Sensor (BIS Quatro Sensor, Medtronic: Catalog No 186-0106) attached on the left side of the head. Values of BIS, total power of electromyographic (EMG) activity, and signal quality index (SQI) were collected via an RS232 port and saved on a laptop computer for off-line analysis.11

After BIS monitoring was initiated, a single trained investigator in each ICU performed the RASS evaluation every 4 hours for 24 hours. Before each RASS assessment, the chief nurse evaluated the SQI of BIS monitoring. The investigator initiated the RASS evaluation as long as the SQI was greater than 75%. The investigator performing the RASS evaluation could not see the monitor during the RASS evaluation. Before and after each RASS evaluation, a 15-minute period of stabilization was implemented to avoid transient changes in the arousal level of the patient because of clinical care, such as change of position, suctioning, radiographic procedures, and other verbal or physical stimulations.

All BIS, EMG, and SQI data were downloaded at 1-minute intervals. During off-line analysis, values of BIS, EMG, and SQI 15 minutes before and after the starting time of RASS evaluation were collected. BIS and EMG values with an SQI ≥75% were included for further analysis. BIS data were collected and analyzed independently by 2 investigators, who were blinded to the result of RASS evaluation.

During the 24-hour period of BIS monitoring and RASS evaluation, we recorded mode of mechanical ventilation, type, and cumulative dose of sedatives used within the previous 4 hours at each RASS evaluation time point.

Back to Top | Article Outline

Definition of BIS Values and Deep Sedation

We used the RASS, a 10-level scale ranging from unarousable (−5) to combative agitation (+4), to evaluate sedation depth (Supplemental Digital Content 1, Table 1, http://links.lww.com/AA/B592).28 Before the study began, the investigators in the 4 participating ICUs were given a 30-minute lecture about the RASS by the principal investigator, and then the technique was demonstrated on 3 patients in the ICU.

We defined the minimum BIS during the 15-minute period before the RASS evaluation as the baseline BIS and the maximum BIS during the 15-minute period after the RASS evaluation as the stimulated BIS. Each RASS evaluation dataset contained the RASS level, the baseline BIS, and the stimulated BIS. In accordance with previous reports and clinical practice guidelines, RASS levels of −3 to −5 were defined as deep sedation.2,3,5,7 Therefore, we divided the datasets into deep sedation and nondeep sedation assessments.

Back to Top | Article Outline

Study Design

The study contained 2 phases. Phase I (training set) was performed in a single center (Daxing Teaching Hospital). Data from these patients were used to determine the threshold value of BIS that best discriminated between datasets with and without the condition of interest (deep sedation). Phase II (validation set) was performed in the 4 participating ICUs. By using the previous selected threshold values in the training set, the diagnostic accuracy of the BIS in detecting deep sedation was thus assessed in the validation set.

Back to Top | Article Outline

Statistical Analysis

In the training set, threshold BIS values for detecting deep sedation were determined by analysis of receiver-operating characteristic (ROC) curves. The area under the ROC curve and 95% confidence intervals (CIs) were reported. The threshold BIS values were identified as those that minimized false negatives (ie, patients who were not assessed as deeply sedated by the BIS but were assessed as deeply sedated by the RASS), with specificity not lower than 50%. This decision was based on the goal of using BIS monitoring as a screening tool for deep sedation, so that the goal of BIS monitoring was to avoid missing the detection of deep sedation. In the validation set, diagnostic accuracy measures of BIS values were characterized by the use of sensitivity, specificity, and predictive values. Positive and negative predictive values also were calculated according to the Bayes’ theorem, using the true prevalence of deep sedation in the validation set.29 The 95% CIs of these characteristics also were calculated.30 The clinical utility index was estimated by the product of sensitivity and positive predictive value.31

For the use of ventilation mode and sedatives between the deep and the nondeep sedation assessment, categorical variables were compared by the Pearson χ2 test or Fisher exact test, and continuous variables by Student t test or Mann-Whitney U test, as appropriate. Repeated measures analysis of variance was used to compare BIS and EMG across time between the deep and the nondeep sedation assessments. Spearman rank-order correlation analysis was used to evaluate the relationship of BIS and EMG with RASS. Paired t test was used to compare the baseline and stimulated BIS and EMG. BIS and EMG were compared across different RASS levels by 1-way analysis of variance followed by Student-Newman-Keuls pairwise comparison.

Statistical analysis was performed with SPSS 17.0 (SPSS Inc, Chicago, IL). A P value of .05 was considered to be statistically significant.

Because we considered the sensitivity of BIS at detecting deep sedation to be the most important feature, the sample size estimation was based on a sensitivity assessment.32 We planned to enroll 42 participants with each having 6 assessments of the RASS (252 assessments in total). With a reported 60% prevalence of deep sedation,2,3,7 we estimated obtaining approximately 150 assessments which would give us a CI precision at the level of 0.14 two-side-width at an expected sensitivity of 85%. In the meantime, we also estimated that 60% of the 252 assessments would yield a CI precision at the level of 0.16 two-side-width if the expected specificity was 50%. By compensating 5% of possible missing data, 270 assessments of RASS evaluation (45 patients) would be required in each of the training set and the validation set.

Back to Top | Article Outline

RESULTS

During the study period, 90 patients were enrolled, with 45 each in the training set (from August to November 2014) and the validation set (from May to June 2015) (Figure 1). Their characteristics are shown in Table 1. Although deep sedation only was prescribed in 6 (7%) patients, at least 1 episode of deep sedation was identified in 76 (84%) patients. During the 24-hour period, 6 RASS assessments were completed in 84 patients, whereas 14 assessments (1 in 2 patients and 3 in 4 patients) were terminated because of low SQI. Thus, a total of 526 assessments were collected, with 262 and 264 in the training set and the validation set, respectively. According to the RASS evaluation, deep sedation occurred in 213 (41%) of 526 assessments (Supplemental Digital Content 2, Figure 1, http://links.lww.com/AA/B593). Characteristics of deep and nondeep sedation assessments are shown in Table 2. The use of controlled mechanical ventilation modes and combined use of propofol and midazolam were more frequent during deep sedation assessments than nondeep sedation assessments. The cumulative dose of propofol during the 4-hour interval before each RASS assessment also was significantly higher during deep sedation assessments.

Table 1.

Table 1.

Table 2.

Table 2.

Figure 1.

Figure 1.

Individual baseline and stimulated BIS and EMG data at each RASS level are presented in Supplemental Digital Content 3, Table 2, http://links.lww.com/AA/B594. In the training set, the incidence of deep sedation was 43% (113 of 262 assessments; Supplemental Digital Content, Figure 1, http://links.lww.com/AA/B593). The areas under the ROC curves (95% CI) for baseline and stimulated BIS were 0.771 (0.714–0.828) and 0.805 (0.752–0.857), respectively (P < .001; Figure 2). The best threshold baseline BIS was 49 (sensitivity of 91.2% and specificity of 51.7%) and stimulated BIS of 82 (sensitivity of 92.0% and specificity of 51.0%; Supplemental Digital Content 4, Table 3, http://links.lww.com/AA/B595). For clinical convenience, the threshold values for identifying deep sedation were modified slightly to ≤50 in baseline BIS (sensitivity of 94.7% and specificity of 49.0%) and ≤80 in stimulated BIS (sensitivity of 85.0% and specificity of 57.0%). In the validation set, the incidence of deep sedation was 38% (100 of 264 assessments; Supplemental Digital Content 2, Figure 1, http://links.lww.com/AA/B593). Using thresholds selected in the training set, diagnostic accuracy measures were calculated in the validation set (Table 3). When baseline BIS and stimulated BIS were combined, the sensitivity, specificity, and clinical utility index were 85.0% (76.1%–91.1%), 85.9% (79.5%–90.7%), and 66.9% (57.8%–76.0%).

Table 3.

Table 3.

Figure 2.

Figure 2.

Figure 3.

Figure 3.

BIS and EMG values differed significantly between deep and nondeep sedation assessments and before and after stimulation (P < .001) (Supplemental Digital Content 5, Figure 2, http://links.lww.com/AA/B596). Both BIS and EMG correlated with RASS measurements (Spearman correlation coefficients ranged from 0.387 to 0.669, P < .001; Figure 3). For either BIS or EMG, the stimulated values were greater than the baseline values (Figure 3, P < .001). Both BIS and EMG values differed across different RASS levels (P < .001). The baseline EMG at RASS = 3 and 4 were greater than other RASS levels (P ranging from .003 to <.001). Stimulated EMG at RASS of −4 and −5 were lower than other RASS levels (P ranging from .003 to <.001).

Back to Top | Article Outline

DISCUSSION

In a population of intubated, sedated ICU patients, we observed a high incidence of “deep” sedation and found that continuous BIS monitoring can reliably detect deep sedation in mechanically ventilated patients. Specifically, a BIS measurement (<50) at baseline and <80.15 minutes later (after “stimulation”) had a 92% sensitivity for detecting a RASS score <−3.

Our findings are consistent with some but not all studies of BIS monitoring of ICU sedation. Although many trials of BIS monitoring in the ICU find poor correlation and no benefit, we found that the BIS monitoring was able to detect deep sedation in mechanically ventilated patients.

One challenge in maintaining a light plane of sedation in the ICU is the difficulty in continuously measuring sedation depth. Serial RASS determinations are labor-intensive because they require a caregiver to test patient behavior. As a result, in resource-poor ICUs, continuous BIS monitoring may reduce the workload of ICU caregivers. In patients with mechanical ventilation, current recommendations suggest that the level of sedation be evaluated at 2- to 4-hour intervals and the sedative dose titrated according to the results of these assessments.5

Nevertheless, despite an increase in sedation evaluation over time,33 deep sedation levels remain common both in observational studies1–4 and in randomized controlled trials targeting light sedation.7 The reason for the persistence of deep sedation is unclear, but the intermittency of sedation scales currently used in clinical practice may play a role.9,10 In addition, sedative accumulation during continuous infusion may predispose to oversedation. The accuracy of sedation depth assessment also may depend on the workload of physicians and nurses.34–36 Finally, when intermittent sedation depth assessments are used, the depth of sedation between measurements remains unknown. These findings suggest that a continuous instrument to monitor sedation depth may have utility in reducing the incidence of inadvertent deep sedation, especially in resource-poor units.

For BIS monitoring in the ICU sedation, another major obstacle is the ICU environment and patient population.11–13,23 These confounders may cause both falsely increased and decreased BIS readings, with EMG activity and electric device interference causing falsely high readings and impaired consciousness and sleep causing falsely low readings.23 To address these possibilities, we excluded brain diseases and pathologic conditions that might impair consciousness. The SQI >70 criteria we used also helped to exclude electric device interference. Although different sedatives can affect the BIS differently, the sedatives used in the present study (propofol, midazolam, dexmedetomidine and fentanyl) had similar impact on BIS values.16,37,38 In our study, therefore, possible contamination of BIS would mostly result from EMG interference and naturally occurring sleep.

Respective electroencephalogram and EMG signals conventionally are located in the 0.5- to 30-Hz band and 30- to 300-Hz band, and the BIS algorithm uses signals in the range of 0.5 to 47 Hz.11 Thus, low-frequency EMG activity (30–47 Hz) may falsely elevate the BIS reading during general anesthesia and deep sedation. Such interference has been considered a major obstacle for the monitoring of BIS in ICU sedation.11–13,23 The A-2000 XP, version 3.x Quattro Sensor measures frontalis muscle EMG activity in the 70- to 110-Hz band, which approximates EMG interference (30–47 Hz) in BIS.

In the present study, we disturbed the patient only minimally during baseline BIS monitoring. At a RASS of −5 to 0, mean baseline EMGs were less than the cutoff point of 35 dB39 in over 90% of our assessments (Figure 3A; Supplemental Digital Content 2, Figure 1, http://links.lww.com/AA/B593). This finding was comparable with previous studies27 and may explain the high NPV in the detection of deep sedation by the baseline BIS (Table 3). After external stimulation during the RASS assessment, both BIS and EMG values increased, and the mean stimulated EMGs were greater than 35 dB at all RASS levels (Figure 3B). This effect suggested that high EMG activity contributed to stimulated BIS values.

Another important confounder of BIS is naturally occurring sleep, which itself may lower the BIS reading.11–13,23,40,41 According to the evaluation procedure, RASS score of −1 or lower could be treated as the patient being in sleep state (Supplemental Digital Content 1, Table 1, http://links.lww.com/AA/B592).28 In our patient population, RASS was evaluated as −1 in 79 of 526 (15%) assessments (Supplemental Digital Content 2, Figure 1, http://links.lww.com/AA/B593). In these 79 patients, the mean baseline BIS was 57 ± 12, and 32% (25/79) assessments were below the threshold value of 50 for identifying deep sedation (Figure 3A; Supplemental Digital Content 3, Table 2, http://links.lww.com/AA/B594). Because these cases had RASS values >−2, they could be considered as not excessively sedated by RASS criteria. However, the incidence of RASS >−3 and BIS <50 was only 5% (25/526) in our overall study cohort. We cannot say whether all patients with that discrepancy were sleeping and clearly do not want to worsen sleep deprivation by disrupting sleeping ICU patients; however, we also note that continuous BIS monitoring may allow clinicians to avoid unnecessary sleep disturbance by identifying patients with baseline BIS, indicating nondeep sedation and RASS = −1. The effect of sedation assessment on sleep merits further study.

Several studies using different target BIS ranges have reported the use of BIS monitoring to augment sedation in ICU patients.22,24–27 In the present study, the primary objective of using BIS was its screening function in deep sedation, and threshold values of BIS were selected in the training set, giving priority to those minimizing false-negative classification with an acceptable specificity. Our results suggest that the use of baseline and stimulated BIS could be combined in clinical practice as a screening tool for possible unwanted deep sedation. As a screening tool, the BIS monitoring has thus several attractive features: objectivity, continuity, high diagnostic power, and easily remembered threshold values (50 and 80). Continuous BIS monitoring may thus have 2 other advantages over intermittent sedation scale evaluation: use as a reminder for performing sedation depth assessment and to allow caregivers to reduce the incidence of sleeping disturbance because of regular RASS assessment.

We took several measures to guarantee data quality. First, the RASS was used as a reference standard for sedation depth assessment, and a single trained investigator in each study location performed the RASS evaluation during the study. The RASS has been validated widely in mechanically ventilated patients and consistently provides a consensus target range.5,9 Recent sedation guidelines also recommended the RASS as a valid and reliable sedation assessment tools in adult ICU patients.5 Second, the RASS evaluator was blinded to the BIS monitor, and the investigators analyzing off-line BIS data were blinded to the results of RASS assessments. Third, we developed threshold values in a single-center training set and tested their diagnostic accuracy in a separate, multicenter validation set. Fourth, we did not deliberately change the local routine practice. Therefore, our results reflect a real clinical scenario.

Our study does have several limitations. First, we only enrolled highly selected group of patients during an early period of mechanical ventilation because of the high reported prevalence of deep sedation at this time and for feasibility of the comparison of the data. Therefore, our results might not be applicable to other populations and clinical situations. Second, in the sample size justification, we estimated that the incidence of deep sedation would be 60%2,3,7; however, the actual incidence was approximately 40% in our study. This discrepancy might have slightly overestimated the statistical power in the present study.

In conclusion, we found that the combined use of baseline BIS with value of 50 and stimulated BIS of 80 was a strong indicator of deep sedation in mechanically ventilated patients. BIS monitoring can be thus be used as an adjunct tool in early screening of undesired deep sedation in a manner different from previous applications of BIS in the ICU. Such an approach may help clinicians at the bedside to reduce undesired episodes or prolonged deep sedation.

Back to Top | Article Outline

DISCLOSURES

Name: Zhu-Heng Wang, MD.

Contribution: This author helped design the study, recruit the patients, collect and analyze the data, and write the manuscript.

Name: Han Chen, MD.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Yan-Lin Yang, MB.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Zhong-Hua Shi, MD.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Qing-Hua Guo, MB.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Yu-Wei Li, MB.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Li-Ping Sun, BS.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Wei Qiao, BS.

Contribution: This author helped recruit the patients and collect and analyze the data.

Name: Guan-Hua Zhou, MD.

Contribution: This author helped design the study and interpret the data.

Name: Rong-Guo Yu, MD.

Contribution: This author helped design the study and interpret the data.

Name: Kai Yin, MD.

Contribution: This author helped design the study and interpret the data.

Name: Xuan He, BS.

Contribution: This author helped analyze and interpret the data.

Name: Ming Xu, BS.

Contribution: This author helped analyze and interpret the data.

Name: Laurent J. Brochard, MD.

Contribution: This author helped design the study, interpret the data, and write the manuscript.

Name: Jian-Xin Zhou, MD.

Contribution: This author helped design the study, analyze and interpret the data, and write the manuscript.

This manuscript was handled by: Avery Tung, MD, FCCM.

Back to Top | Article Outline

ACKNOWLEDGMENTS

We thank Prof Yi-Long Wang (Department of Neurology, Beijing Tiantan Hospital, Capital Medical University) and Prof Hong-Qiu Gu (Clinical Trial and Research Center, Beijing Tiantan Hospital, Capital Medical University) for their valuable suggestions on statistical analysis. We would like to thank Dr Lu Chen (Keenan Research Centre, St Michael’s Hospital, Toronto, Canada) for his help in the revision of the manuscript.

Back to Top | Article Outline

REFERENCES

1. Payen JF, Chanques G, Mantz J, et al. Current practices in sedation and analgesia for mechanically ventilated critically ill patients: a prospective multicenter patient-based study. Anesthesiology. 2007;106:687–695.
2. Shehabi Y, Bellomo R, Reade MC, et al.; Sedation Practice in Intensive Care Evaluation (SPICE) Study Investigators; ANZICS Clinical Trials Group. Early intensive care sedation predicts long-term mortality in ventilated critically ill patients. Am J Respir Crit Care Med. 2012;186:724–731.
3. Shehabi Y, Chan L, Kadiman S, et al.; Sedation Practice in Intensive Care Evaluation (SPICE) Study Group Investigators. Sedation depth and long-term mortality in mechanically ventilated critically ill adults: a prospective longitudinal multicentre cohort study. Intensive Care Med. 2013;39:910–918.
4. Tanaka LM, Azevedo LC, Park M, et al.; ERICC Study Investigators. Early sedation and clinical outcomes of mechanically ventilated patients: a prospective multicenter cohort study. Crit Care. 2014;18:R156.
5. Barr J, Fraser GL, Puntillo K, et al.; American College of Critical Care Medicine. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med. 2013;41:263–306.
6. Treggiari MM, Romand JA, Yanez ND, et al. Randomized trial of light versus deep sedation on mental health after critical illness. Crit Care Med. 2009;37:2527–2534.
7. Shehabi Y, Bellomo R, Reade MC, et al.; Sedation Practice in Intensive Care Evaluation Study Investigators; Australian and New Zealand Intensive Care Society Clinical Trials Group. Early goal-directed sedation versus standard sedation in mechanically ventilated critically ill patients: a pilot study*. Crit Care Med. 2013;41:1983–1991.
8. Hager DN, Dinglas VD, Subhas S, et al. Reducing deep sedation and delirium in acute lung injury patients: a quality improvement project. Crit Care Med. 2013;41:1435–1442.
9. Sessler CN, Grap MJ, Ramsay MA. Evaluating and monitoring analgesia and sedation in the intensive care unit. Crit Care. 2008;12(suppl 3):S2.
10. Tonner PH, Paris A, Scholz J. Monitoring consciousness in intensive care medicine. Best Pract Res Clin Anaesthesiol. 2006;20:191–200.
11. Johansen JW. Update on bispectral index monitoring. Best Pract Res Clin Anaesthesiol. 2006;20:81–99.
12. Fraser GL, Riker RR. Bispectral index monitoring in the intensive care unit provides more signal than noise. Pharmacotherapy. 2005;25:19S–27S.
13. LeBlanc JM, Dasta JF, Kane-Gill SL. Role of the bispectral index in sedation monitoring in the ICU. Ann Pharmacother. 2006;40:490–500.
14. Simmons LE, Riker RR, Prato BS, et al. Assessing sedation during intensive care unit mechanical ventilation with the Bispectral Index and the Sedation-Agitation Scale. Crit Care Med. 1999;27:1499–1504.
15. Riker RR, Fraser GL, Simmons LE, et al. Validating the Sedation-Agitation Scale with the Bispectral Index and Visual Analog Scale in adult ICU patients after cardiac surgery. Intensive Care Med. 2001;27:853–858.
16. Ge SJ, Zhuang XL, Wang YT, et al. Changes in the rapidly extracted auditory evoked potentials index and the bispectral index during sedation induced by propofol or midazolam under epidural block. Br J Anaesth. 2002;89:260–264.
17. Frenzel D, Greim CA, Sommer C, et al. Is the bispectral index appropriate for monitoring the sedation level of mechanically ventilated surgical ICU patients? Intensive Care Med. 2002;28:178–183.
18. de Wit M, Epstein SK. Administration of sedatives and level of sedation: comparative evaluation via the Sedation-Agitation Scale and the Bispectral Index. Am J Crit Care. 2003;12:343–348.
19. Lu CH, Ou-Yang HY, Man KM, et al. Relative reliability of the auditory evoked potential and Bispectral Index for monitoring sedation level in surgical intensive care patients. Anaesth Intensive Care. 2008;36:553–559.
20. Trouiller P, Fangio P, Paugam-Burtz C, et al. Frequency and clinical impact of preserved bispectral index activity during deep sedation in mechanically ventilated ICU patients. Intensive Care Med. 2009;35:2096–2104.
21. Yaman F, Ozcan N, Ozcan A, et al. Assessment of correlation between bispectral index and four common sedation scales used in mechanically ventilated patients in ICU. Eur Rev Med Pharmacol Sci. 2012;16:660–666.
22. Karamchandani K, Rewari V, Trikha A, et al. Bispectral index correlates well with Richmond agitation sedation scale in mechanically ventilated critically ill patients. J Anesth. 2010;24:394–398.
23. Dahaba AA. Different conditions that could result in the bispectral index indicating an incorrect hypnotic state. Anesth Analg. 2005;101:765–773.
24. Memiş D, Turan A, Karamanlioglu B, et al. Comparison of sufentanil with sufentanil plus magnesium sulphate for sedation in the intensive care unit using bispectral index. Crit Care. 2003;7:R123–R128.
25. Weatherburn C, Endacott R, Tynan P, et al. The impact of bispectral index monitoring on sedation administration in mechanically ventilated patients. Anaesth Intensive Care. 2007;35:204–208.
26. Olson DM, Zomorodi MG, James ML, et al. Exploring the impact of augmenting sedation assessment with physiologic monitors. Aust Crit Care. 2014;27:145–150.
27. Dahaba AA, Lischnig U, Kronthaler R, et al. Bispectral-index-guided versus clinically guided remifentanil/propofol analgesia/sedation for interventional radiological procedures: an observer-blinded randomized study. Anesth Analg. 2006;103:378–384.
28. Sessler CN, Gosnell MS, Grap MJ, et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med. 2002;166:1338–1344.
29. Bayes T. An essay towards solving a problem in the doctrine of chances. 1763. MD Comput. 1991;8:157–171.
30. Harper R, Reeves B. Reporting of precision of estimates for diagnostic accuracy: a review. BMJ. 1999;318:1322–1323.
31. Mitchell AJ. Sensitivity × PPV is a recognized test called the clinical utility index (CUI+). Eur J Epidemiol. 2011;26:251–252.
32. Obuchowski NA, McClish DK. Sample size determination for diagnostic accuracy studies involving binormal ROC curve indices. Stat Med. 1997;16:1529–1542.
33. Sneyers B, Laterre PF, Perreault MM, et al. Current practices and barriers impairing physicians’ and nurses’ adherence to analgo-sedation recommendations in the intensive care unit—a national survey. Crit Care. 2014;18:655.
34. Gill KV, Voils SA, Chenault GA, et al. Perceived versus actual sedation practices in adult intensive care unit patients receiving mechanical ventilation. Ann Pharmacother. 2012;46:1331–1339.
35. Sneyers B, Laterre PF, Bricq E, et al. What stops us from following sedation recommendations in intensive care units? A multicentric qualitative study. J Crit Care. 2014;29:291–297.
36. Tingsvik C, Bexell E, Andersson AC, et al. Meeting the challenge: ICU-nurses’ experiences of lightly sedated patients. Aust Crit Care. 2013;26:124–129.
37. Yang KS, Habib AS, Lu M, et al. A prospective evaluation of the incidence of adverse events in nurse-administered moderate sedation guided by sedation scores or Bispectral Index. Anesth Analg. 2014;119:43–48.
38. Xu J, Jin C, Cui X, et al. Comparison of dexmedetomidine versus propofol for sedation after uvulopalatopharyngoplasty. Med Sci Monit. 2015;21:2125–2133.
39. Mathews DM, Kumaran KR, Neuman GG. Bispectral index-derived facial electromyography-guided fentanyl titration in the opiate-exposed patient. Anesth Analg. 2003;96:1062–1064.
40. Tung A, Lynch JP, Roizen MF. Use of the BIS monitor to detect onset of naturally occurring sleep. J Clin Monit Comput. 2002;17:37–42.
41. Dahaba AA, Xue JX, Xu GX, et al. Bilateral Bispectral Index (BIS)-Vista as a measure of physiologic sleep in sleep-deprived anesthesiologists. Minerva Anestesiol. 2011;77:388–393.

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2016 International Anesthesia Research Society